Sparse regression with multi-type regularized feature modeling
From MaRDI portal
Publication:2657005
DOI10.1016/J.INSMATHECO.2020.11.010zbMATH Open1460.91218arXiv1810.03136OpenAlexW2895600735MaRDI QIDQ2657005FDOQ2657005
Authors: Sander Devriendt, Katrien Antonio, Tom Reynkens, Roel Verbelen 
Publication date: 17 March 2021
Published in: Insurance Mathematics \& Economics (Search for Journal in Brave)
Abstract: Within the statistical and machine learning literature, regularization techniques are often used to construct sparse (predictive) models. Most regularization strategies only work for data where all predictors are treated identically, such as Lasso regression for (continuous) predictors treated as linear effects. However, many predictive problems involve different types of predictors and require a tailored regularization term. We propose a multi-type Lasso penalty that acts on the objective function as a sum of subpenalties, one for each type of predictor. As such, we allow for predictor selection and level fusion within a predictor in a data-driven way, simultaneous with the parameter estimation process. We develop a new estimation strategy for convex predictive models with this multi-type penalty. Using the theory of proximal operators, our estimation procedure is computationally efficient, partitioning the overall optimization problem into easier to solve subproblems, specific for each predictor type and its associated penalty. Earlier research applies approximations to non-differentiable penalties to solve the optimization problem. The proposed SMuRF algorithm removes the need for approximations and achieves a higher accuracy and computational efficiency. This is demonstrated with an extensive simulation study and the analysis of a case-study on insurance pricing analytics.
Full work available at URL: https://arxiv.org/abs/1810.03136
Recommendations
- Adaptive regularization using the entire solution surface
- Smoothing proximal gradient method for general structured sparse regression
- Sparsity and Smoothness Via the Fused Lasso
- Variable selection in multivariate linear models for functional data via sparse regularization
- A sparse regularization approach with Log type penalty
Cites Work
- A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems
- Simultaneous Factor Selection and Collapsing Levels in ANOVA
- Estimating the dimension of a model
- The Adaptive Lasso and Its Oracle Properties
- Flexible smoothing with \(B\)-splines and penalties. With comments and a rejoinder by the authors
- Coherent dispersion criteria for optimal experimental design
- Least angle regression. (With discussion)
- Title not available (Why is that?)
- Title not available (Why is that?)
- A new look at the statistical model identification
- A dual algorithm for the solution of nonlinear variational problems via finite element approximation
- Title not available (Why is that?)
- Sparsity and Smoothness Via the Fused Lasso
- Title not available (Why is that?)
- Regularization and Variable Selection Via the Elastic Net
- Sparse modeling of categorial explanatory variables
- Model Selection and Estimation in Regression with Grouped Variables
- The solution path of the generalized lasso
- Properties and refinements of the fused Lasso
- Generalized additive models
- A uniform framework for the combination of penalties in generalized structured models
- Title not available (Why is that?)
- Title not available (Why is that?)
- Relaxed Lasso
- A note on adaptive group Lasso
- Non-life rate-making with Bayesian GAMs
- Nonlife ratemaking and risk management with Bayesian generalized additive models for location, scale, and shape
- Restricted Estimation of Generalized Linear Models
- On the robustness of the generalized fused Lasso to prior specifications
- A data driven binning strategy for the construction of insurance tariff classes
Cited In (11)
- A non-convex regularization approach for stable estimation of loss development factors
- Generalized fused Lasso for grouped data in generalized linear models
- REC: fast sparse regression-based multicategory classification
- Insurance pricing with hierarchically structured data an illustration with a workers' compensation insurance portfolio
- Identifying the determinants of lapse rates in life insurance: an automated Lasso approach
- Multi-state modelling of customer churn
- Variable Selection Using a Smooth Information Criterion for Distributional Regression Models
- Mixture Composite Regression Models with Multi-type Feature Selection
- Efficient path algorithms for clustered Lasso and OSCAR
- Loss amount prediction from textual data using a double GLM with shrinkage and selection
- A group regularisation approach for constructing generalised age-period-cohort mortality projection models
Uses Software
This page was built for publication: Sparse regression with multi-type regularized feature modeling
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2657005)
